基于美国FDA不良事件报告系统数据库的免疫检查点抑制剂致糖尿病风险分析

Analysis on the risk of diabetes mellitus related to immune checkpoint inhibitors based on the US FDA Adverse Event Reporting System

目的分析免疫检查点抑制剂(ICI)致糖尿病的风险。方法收集美国FDA不良事件报告系统2004年第1季度至2021年第2季度收录的度伐利尤单抗、帕博利珠单抗、纳武利尤单抗和阿替利珠单抗致糖尿病[包括暴发性1型糖尿病(FT1DM)、1型糖尿病(T1DM)、糖尿病酮症酸中毒(DKA)]不良事件(AE)报告,采用比例报告比值比(PRR)法评价4种药物与FT1DM、T1DM、DKA间的相关性。AE报告数≥3、PRR≥2且χ^(2)≥4的AE被判定为与药物有统计学相关性,PRR越大,AE与药物的相关性越强,风险信号越强。结果收集到的度伐利尤单抗、帕博利珠单抗、纳武利尤单抗和阿替利珠单抗相关糖尿病AE报告分别为53、386、957和72例,共1468例。度伐利尤单抗、帕博利珠单抗、纳武利尤单抗和阿替利珠单抗与FT1DM相关性的PRR分别为21.97(χ^(2)=40.71)、71.50(χ^(2)=3531.21)、294.30(χ^(2)=43915.75)和33.58(χ^(2)=279.70);与T1DM相关性的PRR分别为12.12(χ^(2)=162.08)、21.04(χ^(2)=3391.17)、20.99(χ^(2)=5816.11)和9.71(χ^(2)=224.81);与DKA相关性的PRR分别为6.93(χ^(2)=161.26)、4.78(χ^(2)=426.52)、6.82(χ^(2)=1797.15)和3.04(χ^(2)=41.84)。4种药物与目标AE均有统计学相关性,致目标AE的风险信号强度均为FT1DM>T1DM>DKA;致FT1DM的风险信号强度为纳武利尤单抗>帕博利珠单抗>阿替利珠单抗>度伐利尤单抗,致T1DM的风险信号强度为帕博利珠单抗≈纳武利尤单抗>度伐利尤单抗>阿替利珠单抗;致DKA的风险信号强度为度伐利尤单抗≈纳武利尤单抗>帕博利珠单抗>阿替利珠单抗。结论度伐利尤单抗、帕博利珠单抗、纳武利尤单抗和阿替利珠单抗均可导致糖尿病,以致FT1DM的风险信号最强,T1DM次之,DKA较弱。

Objective To analyze the risk of diabetes mellitus related to immune checkpoint inhibitors(ICI).Methods The adverse event(AE)reports on fulminant type 1 diabetes mellitus(FT1DM),type 1 diabetes mellitus(T1DM),diabetic ketoacidosis(DKA),which were related to duvalizumab,pabolizumab,nivolumab,and atezolizumab in the US FDA Adverse Event Reporting System from the first quarter of 2004 to the second quarter of 2021,were collected.The correlation between the 4 drugs and FT1DM,T1DM,and DKA were evaluated using proportional reporting odds ratio(PRR)method.AE with reports≥3,PRR value≥2,andχ^(2)≥4 were judged to have statistical correlations with drugs.The greater the PRR value,the stronger the correlation between AE and drugs and the stronger the risk signals.Results A total of 1468 AE reports on diabetes mellitus were collected,53,386,957,and 72 of which were related toduvalizumab,pabolizumab,nivolumab,and atezolizumab,respectively.For duvalizumab,pabolizumab,nivolumab,and atezolizumab,the PRR reflecting the correlation with FT1DM were 21.97(χ^(2)=40.71),71.50(χ^(2)=3531.21),294.30(χ^(2)=43915.75),and 33.58(χ^(2)=279.70),respectively;the PRR reflecting correlation with T1DM were 12.12(χ^(2)=162.08),21.04(χ^(2)=3391.17),20.99(χ^(2)=5816.11),and 9.71(χ^(2)=224.81),respec‑tively;the PRR reflecting correlation with DKA were 6.93(χ^(2)=161.26),4.78(χ^(2)=426.52),6.82(χ^(2)=1797.15),and 3.04(χ^(2)=41.84),respectively.The 4 drugs were statistically correlated with their corresponding AE.The order of risk signal intensity for corresponding AE was FT1DM>T1DM>DKA.The order of risk signal intensity for FT1DM were nivolumab>pabolizumab>duvalizumab>atezolizumab,for T1DM were pabolizumab≈nivolumab>duvalizumab>atezolizumab,for DKA were duvalizumab≈nivolumab>pabolizumab>atezolizumab.Conclusions Duvalizumab,pabolizumab,nivolumab,and atezolizumab all can cause diabetes mellitus.The risk signal intensity was the strongest for FT1DM,followed by T1DM and DKA in order.